Senior AI engineer - Agent system | AI platform

Compensation: $80k – $120k • No equity

Build the platform behind production AI agents.

We’re looking for a Senior AI Engineer to design and build the runtime, orchestration, retrieval, evaluation, and platform infrastructure behind production AI agents.

This is not a prompt-engineering or chatbot-integration role. We’re looking for a strong software engineer who can take agent systems beyond prototypes and make them reliable, observable, and safe enough to operate in production.

What You’ll Own

• Design and build agent runtimes for multi-step, stateful workflows • Build orchestration for tool calling, retries, approvals, human-in-the-loop flows, and failure recovery • Develop shared platform services for model access, routing, structured outputs, and context management • Build retrieval and RAG systems using embeddings, vector search, hybrid retrieval, and reranking • Create evaluation and regression systems for agent quality and release safety • Build observability around traces, latency, failures, token usage, and cost • Design APIs and SDKs that allow product teams to build on the platform without reimplementing agent infrastructure • Own systems from architecture and implementation through deployment and production operation

What We’re Looking For

• 8+ years of software engineering experience with strong backend or distributed-systems fundamentals • Strong Python, Go, or TypeScript experience • Hands-on experience shipping LLM, agent, or AI-platform systems to production
• Experience with APIs, asynchronous systems, stateful workflows, and production reliability • Experience with cloud-native systems such as AWS, Kubernetes, Docker, PostgreSQL, Kafka, or Redis

Nice to Have

• Multi-agent or agent-runtime architecture • LLM evaluation, trace replay, or LLM-as-judge systems • MCP, model gateways, model routing, or shared AI developer platforms • RAG, vector databases, hybrid retrieval, and reranking • OpenTelemetry, Terraform, or production AI observability

How You’ll Work

You’ll work with a small engineering team with significant ownership and relatively little process overhead. You should be comfortable making architectural tradeoffs, shipping incrementally, debugging production behavior, and deciding when to build infrastructure versus adopting existing tools.

Role

Remote · Contractor 50-80/hr

About us https://www.orbion.bond/